From the Society for Clinical Vascular Surgery

Divergent outcomes after percutaneous therapyfor symptomatic renal artery stenosisNayan Sivamurthy, MD,a Scott M. Surowiec, MD,a Eva Culakova, PhD,b Jeffrey M. Rhodes, MD,a

David Lee, MD,c Yaron Sternbach, MD,a David L. Waldman, MD,c Richard M. Green, MD,a and MarkG. Davies, MD, PhD,a,c Rochester, NY

Objective: Percutaneous intervention for symptomatic renal artery atherosclerosis is rapidly replacing surgery in manycenters. This study evaluated the anatomic and functional outcomes of endovascular therapy for atherosclerotic renalartery stenosis on a combined vascular surgery and interventional radiology service at an academic medical center.Methods: This was a retrospective analysis of patients who underwent renal artery angioplasty with or without stentingbetween January 1990 and June 2002. Indications included hypertension (86%) and rising serum creatinine concentra-tion (55%). One hundred forty-six patients (80 women; average age, 71 years [range, 44-89 years]) underwent 183attempted interventions (64 to treat bilateral stenosis). Forty-five percent of patients had significant bilateral disease: 27%had greater than 50% bilateral stenosis, and the remainder had nonfunctioning, absent, or occluded vessels.Results: Of 183 planned interventions, technical success (<30% residual stenosis) was achieved in 179 vessels (98%) withplacement of 137 stents (75%). Thirty-day mortality was 0.7%. The major morbidity rate was 4%, and the procedure-related complication rate was 18%. Five-year cumulative patient mortality was 25%. Primary patency, assisted primarypatency, and recurrent stenosis rates were 82% � 9%, 100% � 0%, and 30% � 7%, respectively, at 5 years. Within 3 monthsof the procedure, 52% of patients who received treatment of hypertension demonstrated clinical benefit (hypertensionimproved or cured), which was maintained in 68% of patients at 5 years. Serum creatinine concentration was lowered orstabilized in 87% of patients within 3 months of the procedure, but this benefit, including freedom from dialysis, wasmaintained in only 45% of patients at 5 years.Conclusions: Endovascular intervention for symptomatic atherosclerotic renal artery stenosis is technically successful.There were excellent patency and low recurrent stenosis rates. There is immediate clinical benefit for most patients, butdivergent long-term functional outcomes. Endovascular interventions modestly enhance the care of the patient withhypertension, but poorly preserve long-term renal function in the patient with chronic renal impairment. (J Vasc Surg2004;39:565-74.)

Atherosclerotic renal artery stenosis is a significantcause of poorly controlled hypertension and progressiverenal dysfunction, leading to ischemic nephropathy andother end organ damage.1 Renal artery stenosis progressesin 51% of renal arteries within 5 years of diagnosis.2 Asmany as 5% of renal arteries with greater than 60% stenosisper year will eventually occlude, with resultant renal atro-phy.3 Recently the attention paid to renal artery disease hasintensified with the greater use and availability of catheter-based interventions.

Historically, surgical revascularization was the acceptedstandard of therapy, but was limited by relatively highprocedural morbidity and mortality.4 Catheter-based treat-ment is an attractive alternative to surgical revasculariza-tion, because it is tolerated by a greater proportion of

From the Center for Vascular Disease, Department of Surgery, Division ofVascular Surgerya, Department of Biostatisticsb, and Department of Ra-diology, Division of Interventional Radiology,c University of Rochester.

Competition of interest: none.Presented at the Thirty-first Annual Meeting of the Society for Clinical

Vascular Surgery, Miami, Fla, Mar 4-9, 2003.Reprint requests: Mark G. Davies, MD, PhD, Center for Vascular Disease,

Division of Vascular Surgery, Department of Surgery, University of Roch-ester Medical Center, 601 Elmwood Ave, Box SURG, Rochester, NY14642 (e-mail: mark_davies@urmc.rochester.edu).

0741-5214/$30.00Copyright © 2004 by The Society for Vascular Surgery.doi:10.1016/j.jvs.2003.09.024

affected patients. Recent studies have shown the immediateand long-term technical success of the procedure.5-8 How-ever, the optimal treatment of renovascular disease contrib-uting to hypertension and renal dysfunction is not known.Furthermore, the long-term functional outcomes of en-doluminal therapy on hypertension and renal preservationremain controversial.6,9,10 This study reports the anatomicand functional outcomes of endovascular therapy for ath-erosclerotic renal artery stenosis in a clinically and demo-graphically mixed patient population on a combined vascu-lar surgery and interventional radiology service at anacademic medical center.

METHODS

Experimental design and setting. We performed aretrospective review of inpatient and office records forpatients who underwent primary renal artery angioplasty orstenting to treat atherosclerotic renal artery disease at theUniversity of Rochester Medical Center between January1990 and June 2002. For each patient, demographic data,existing comorbid conditions, and risk factors for athero-sclerosis were identified. Median patient follow-up was 1.3years (average, 2.1 � 2.3 years; range, 0-10 years). Eighty-nine percent of patients underwent duplex ultrasound scan-ning at last follow-up. Last average follow-up was 1.8 � 2.0years (range, 0-10 years).

565

JOURNAL OF VASCULAR SURGERYMarch 2004566 Sivamurthy et al

Treatment algorithm. Patients with hypertension, el-evated serum creatinine concentration, congestive heartfailure, or diminishing renal mass underwent a diagnosticstudy to identify the presence of renal artery stenosis. Thisstudy consisted of standard angiography, magnetic reso-nance angiography, renal isotope scanning, or duplex ultra-sound scanning. Duplex ultrasound scanning criteria usedto identify renal artery stenosis have been described.3 In thepresence of clinical criteria defined by Rundback et al11 andstenosis of 60% or greater on ultrasound scanning or mag-netic resonance angiography, or a positive renal scan, an-giography was performed. Interventions were performed inpatients with renal artery stenosis 50% or greater at angiog-raphy, regardless of comorbid conditions. Patients who didnot meet the clinical criteria referenced were managedmedically. Nonfunctioning kidneys were not treated. In thepresence of bilateral disease, the clinical criteria for treat-ment of both kidneys were the same as those for solitaryrenal artery stenosis. Patients were followed up at 6-monthintervals after the procedure. Blood pressure,11 serum cre-atinine concentration, and number of antihypertensivemedications were identified at these follow-up intervals. Byfollowing the methods by Rundback et al,11 reliability ofblood pressure was ensured and observer bias was elimi-nated. Duplex ultrasound scanning was performed in ourvascular laboratory, which is registered by the IntersocietalCommission for Accredication of Vascular Laboratories.Each patient underwent duplex ultrasound scanning atleast once within 6 months of the procedure, and mostunderwent ultrasound at last follow-up to assess patency. Ifthe duplex ultrasound scan showed stenosis of 60% orgreater and the patient had recurrent symptoms, angiogra-phy was performed and recurrent stenosis was treated if50% or greater.

Definitions. Hypertension was defined as systolicblood pressure greater than 140 mm Hg or diastolic bloodpressure greater than 90 mmHg on three occasions over 6months. An elevated creatinine concentration was definedas 1.5 mg/dL or greater on two consecutive measurementsover 3 months. Chronic renal insufficiency was defined asserum creatinine concentration persistently greater than1.5 mg/dL for more than 6 months. The baseline serumcreatinine concentration was the value recorded closest tothe procedure. Patients were considered to have a “non-functioning kidney” if any two of the following local crite-ria used at our institution over the time of the study weremet: (1) a duplex ultrasound scan demonstrated a pole-to-pole length of less than 9 cm, with no renal flow in the mainrenal artery and parenchymal peak systolic velocity less than10 cm/s; (2) surgically or congenitally absent kidney; (3)no visible nephrogram at contrast material–enhanced arte-riography. Ostial lesions were defined as lesions within 1 cmof the renal artery orifice. The procedure was considered asuccess if residual stenosis was less than 30%, and a failure ifresidual stenosis was 30% or greater, according to angio-graphic measurement, including lesions that could not bedilated or crossed. Death within 30 days of the procedurewas considered procedure-related. Primary patency was

defined as absence of occlusion or interventions to treatrecurrent stenosis of the vessel after the primary interven-tion. Assisted primary patency was defined as absence ofocclusion of the vessel after the primary intervention, butincluding any interventions to treat recurrent stenosis. Pri-mary recurrent stenosis was defined as a treated segmentwith a duplex classification of stenosis 60% or greater.Secondary recurrent stenosis was defined as a treated seg-ment with a duplex classification of stenosis 60% or greaterafter additional percutaneous interventions.12 Response inpatients with hypertension was defined as follows11:“cured” hypertension was normotension, that is, diastolicblood pressure less than 90 mm Hg and systolic bloodpressure less than 140 mm Hg, without medications; “im-proved” hypertension was normotension, that is, diastolicblood pressure less than 90 mm Hg and/or systolic bloodpressure less than 140 mm Hg, with the same or fewermedications, or diastolic blood pressure 15 mm Hg belowbaseline with the same or fewer medications; “no effect”was no change or inability to meet these criteria for cure orimprovement of hypertension, and was considered a treat-ment failure. Recurrence of hypertension was defined asrecurrence of diastolic blood pressure greater than 90 mmHg or systolic blood pressure greater than 140 mm Hg ora greater than 20% increase in post-procedure baselinediastolic blood pressure that was not treatable with repeatangioplasty. Early and late renal function responses toangioplasty were defined as follows: “cured” renal functionrequired a serum creatinine concentration less than 1.5mg/dL; “improvement” in renal function required agreater than 20% reduction in the serum creatinine concen-tration; “deterioration” in renal function was considered atreatment failure, and required a greater than 20% increasein the serum creatinine concentration, institution of dialy-sis, or renal related mortality.12 Patients who had a 20% orless change in serum creatinine concentration after theprocedure were considered to have “stable” renal function.

Statistical analysis. We performed our analysis on an“intention to treat” basis. Measured values are reported aspercentage or mean � 1 SD. Survival, cumulative primarypatency, primary assisted patency, recurrent stenosis, andclinical benefit rates were calculated with life table analysis,and are reported with American Association for VascularSurgery/Society for Vascular Surgery criteria. Standarderrors are reported in life table analyses. The log-rank testwas used to determine differences between life tables. Anal-yses were performed with SAS software (version 8.2; SASInstitute, Cary NC). Multivariate stepwise regression anal-ysis was used to determine the influence of preproceduraland periprocedural factors on outcomes. The significancelevel P � .10 was used to include or eliminate a covariatefrom the model. Covariates were considered significantlyassociated with the outcome if they were included in thefinal model and their significance level was P � .05. Inter-actions between statistically significant covariates werechecked. The Cox proportional hazards model was used fortime-dependent outcomes (mortality, recurrent stenosis,long-term freedom from worsening hypertension, long-

JOURNAL OF VASCULAR SURGERYVolume 39, Number 3 Sivamurthy et al 567

term freedom from renal related morbidity and mortality).The proportional hazards assumption was checked andconfirmed for the significant covariates with log (�log[survival]) curves. Association of mortality with recurrentstenosis, need for hemodialysis, and long-term clinical out-comes was checked in separate univariate models with thetime-dependent covariate. Logistic regression models wereused for outcomes measured shortly after the procedure(complications, short-term clinical benefits). The depen-dence of each covariate on the outcome was first checkedseparately with the �2 test. Covariates with a significancelevel of P � .10 were included in the multivariate stepwiseanalysis.

RESULTS

Patient population. One hundred forty-six patientswere identified (66 men, 80 women; mean age, 71 � 10years [range, 44-89 years]). Comorbid conditions and ath-erosclerotic risk factors were typical for a population withsymptomatic peripheral vascular disease (Table I). Onehundred eighty-three stenotic renal arteries present in 146patients required intervention. Principal indications for in-tervention were hypertension (86%), elevated creatinineconcentration (55%), and hypertension and elevated creat-inine concentration (42%; Table I). No patients withouthypertension and antihypertensive medications receivedtreatment. Diagnostic imaging, used to define the stenosisand the anatomic and functional findings in each kidney,are reviewed in Tables II and III. There was significantvascular disease in both kidneys in 45% of patients (Table

Table I. Patient characteristics and indications forintervention

Characteristic n %

Total patients 146Male 66 45Female 80 55

CardiacMyocardial infarction 53 36Congestive heart failure 44 30Coronary artery disease 74 50Hypertension 145 99

Diabetes mellitusInsulin-dependent 19 13Non-insulin-dependent 24 16

Transient ischemic attack or stroke 36 25Chronic renal insufficiency 51 35Chronic obstructive pulmonary disease 19 13Coagulopathy 3 2Hyperlipidemia 76 52Smoking 69 47Previous vascular surgery 43 29Indications

Hypertension 126 86Elevated creatinine concentration 80 55Hypertension and elevated

creatinine concentration62 42

Pulmonary edema 13 9Shrinking kidney mass 3 2Asymptomatic disease 5 3

II). Specific data related to each intervention and subse-quent procedures in solitary and bilateral diseased kidneysare summarized in Table III. One hundred fifteen of 137stented lesions were ostial (84%). Thirty two of 39 patients(82%) with angiographically confirmed bilateral renal arterystenosis (�50%) received treatment to obtain maximaltherapeutic benefit (64 of 78 arteries). Renal artery occlu-sion in patients with bilateral disease was confirmed in all 15patients at angiography. In the remainder of patients thecontralateral kidney was not addressed, because the kidneywas nonfunctioning (n � 1), the patient died before theopposite side could be addressed, or the patient refusedtherapy.

Immediate outcomes (<3 months). In 4 vessels(2.2%) the lesion could not be crossed, and these weredeemed immediate intention-to-treat failures. Technicalsuccess was achieved in 179 vessels (98%). Post-procedurevessel diameter was 5.5 � 0.8 mm (range, 3-8 mm).Thirty-three procedure-related complications occurred in27 interventions (18%); 9 of these were related to the accesssite (5%). These included 7 stent dislodgements, 7 bleedingepisodes, 5 groin hematomas, 4 acute renal artery dissec-tions, 4 femoral or axillary pseudoaneurysms, 3 proceduralkidney perforations, 2 peripheral embolizations to the foot,and 1 renal artery occlusion. Four patients required surgerysecondary to procedural complications (3 explorations be-cause of subcapsular or retroperitoneal hematomas, 1 sur-gical revascularization). These complications did not resultin the loss of any kidney. Female gender and elevatedpre-procedure diastolic blood pressure were associated withincreased morbidity (Table IV). Major morbidity was 4%.One major adverse cardiac event occurred after the proce-dure, resulting in a 30-day mortality and in-hospital mor-tality of 0.7%. No deaths occurred within 90 days as a resultof renal artery injury.

Immediate clinical benefit for patients with hyperten-sion and renal impairment within 3 months of the proce-dure is summarized in Table V. Fifty-two percent of pa-tients who received treatment of hypertension showedimmediate clinical benefit (improved or cured hyperten-sion). In patients who received treatment of elevated serumcreatinine concentration (�1.5 mg/dL), 87% showed sta-bility of or decrease in serum creatinine concentration afterthe procedure. Diastolic blood pressure was statisticallyimproved within 3 months after the procedure in patientsreceiving treatment of hypertension, elevated creatinineconcentration, or both. Systolic blood pressure was statis-tically improved within 3 months after the procedure inpatients receiving treatment of hypertension. Patients withcerebrovascular disease and elevated pre-procedure systolicblood pressure demonstrated immediate treatment successfor hypertension. Pre-procedure creatinine concentrationgreater than 2.5 mg/dL was associated with deteriorationof creatinine concentration within 3 months of the proce-dure (Table IV).

Outcomes (>3 months). Patient mortality for total,hypertensive, and renal dysfunction population by life tableanalysis was 25%, 21%, and 27% at 5 years (Fig). Nineteen

JOURNAL OF VASCULAR SURGERYMarch 2004568 Sivamurthy et al

patients died by last follow-up; 8 deaths were cardiac re-lated, 6 were not cardiac related, and 5 were renal related.Immediate and long-term creatinine concentration deteri-oration and dialysis post-procedure were associated withincreased mortality. Elevated pre-procedure systolic blood

Table II. Preprocedural data

Bilateral vessel statusStenosis (�50%) � normalStenosis (�50) � stenosis (�50%)Bilateral Stenosis (�50%)Stenosis (�50%) � occlusionStenosis (�50%) � absent

HTN

Ipsilateral kidneySize (cm) 10.4 � 1.3Resistive index 0.73 � 0.10

Contralateral kidneySize (cm) 9.9 � 1.6*Resistive index 0.74 � 0.13

Creatinine concentration (mg/dL) 1.7 � 1.2*Creatinine clearance (mL/min) 48 � 27*Systolic blood pressure (mm Hg) 181 � 28Diastolic blood pressure (mm Hg) 90 � 17Number of medications 3.0 � 1.0

HTN, Hypertension; Cr, creatinine concentration.*P � .05.

Table III. Diagnostic, anatomic, and procedural data

n %

Procedures 183Diagnostic imaging

Angiography 176 98Duplex ultrasound scanning 47 26Magnetic resonance angiography 32 18Isotope scanning 30 17

SideLeft 100 55Right 83 45

Number of renal vessels (mean � SD) 1.1 � 0.2Number of lesions (mean � SD) 1.0 � 0.1Degree of vessel stenosis (%) 80 � 10Site of stenosis

Ostial 138 75Proximal 39 21Middle 6 3

ProcedureAngioplasty 46 25Primary stent placement 125 68Total stent placement 137 75

Contrast mediumIonic 166 91Gadolinium or carbon dioxide 17 9

Subsequent ipsilateral procedure 13 7.1Bilateral kidney treatment 64* 35Complications 33 18

*Sixty-four arteries in 32 patients.

pressure was associated with lower mortality. Recurrentstenosis and treatment failure for hypertension were notassociated with mortality (Table IV). When the populationwas stratified by pre-procedure creatinine concentration, inthose with a creatinine concentration of 2.5 mg/dL orgreater the mortality rate was 38%, whereas in those with acreatinine concentration less than 2.5 mg/dL the mortalityrate was 23%, at 5 years (P � .08). Analysis by creatinineclearance was performed, and yielded similar results as withserum creatinine concentration, and gave no additionalimprovement in predictive ability.

The primary patency rate was 82% � 9% at 5 years.There were 17 endoluminal interventions in 13 patients totreat recurrent stenosis. There was only one vessel occlu-sion at 5.5 years. The 5-year rate of repeat intervention was14% � 6% (95% confidence interval [CI], 3%-25%), with amedian time of 0.7 years to this procedure. After repeatinterventions to treat recurrent stenosis the primary assistedpatency rate was 100% at 5 years (Fig). By multivariateanalysis, secondary interventions did not influence the pri-mary end points of this study. The primary recurrent ste-nosis rate was 30% � 7% at 5 years, and the secondaryrecurrent stenosis rate was 0% at 5 years. Five surgicalrevascularization procedures to treat recurrent symptomswith recurrent stenosis were required in 4 patients treatedpercutaneously during follow-up of more than 5 years. Nofactors were associated with the development of recurrentstenosis.

Cumulative freedom from worsening hypertension was89% � 3%, 78% � 6%, and 68% � 9% at 1, 3, and 5 years,respectively (Fig). Hyperlipidemia was associated with free-dom from worsening hypertension. Cumulative freedomfrom dialysis, renal disease–related mortality, or greater

No. ofpatients %

68 4712 839 2715 1012 8

Cr Total Range

10.2 � 1.2 10.3 � 1.2 7.0–13.70.75 � 0.10 0.73 � 0.10 0.32–0.89

9.2 � 1.6* 9.9 � 1.5 4.7–13.00.73 � 0.20 0.74 � 0.13 0.45–0.932.4 � 1.3* 1.8 � 1.2 0.6–9.133 � 20* 46 � 26 3–163

168 � 28 176 � 30 100–24087 � 18 88 � 18 40–1403.0 � 1.0 2.9 � 1.1 0.0–6.0

JOURNAL OF VASCULAR SURGERYVolume 39, Number 3 Sivamurthy et al 569

than 20% elevation in post-procedure serum creatinineconcentration was 67% � 5%, 50% � 7%, and 45% � 9% at1, 3, and 5 years, respectively. When stratified by pre-procedure serum creatinine concentration of less than orgreater than 2.5 mg/dL, this cumulative freedom was 39%� 8% (95% CI, 23%-55%) and 30% � 12% (95% CI,7%-53%) at 5 years, respectively (P � .02). Analysis accord-ing to creatinine clearance was performed, and yieldedsimilar results as those with serum creatinine concentration,and gave no additional improvement in predictive ability.In patients with creatinine concentration less than 1.5mg/dL (n � 66), 97% within 3 months and 72% within 5years had a cumulative freedom from dialysis, renal disease–related mortality, or greater than 20% elevation in post-procedure serum creatinine concentration. Immediatetreatment success for creatinine was associated with in-creased freedom from dialysis, renal disease–related mortal-ity, or greater than 20% elevation in post-procedure serumcreatinine concentration (Table IV). Immediate treatmentsuccess for hypertension and creatinine was associated withreduced renal- related morbidity and mortality for all pa-tients.

Twenty three patients ultimately were dialysis-depen-dent within 3 years of their interventions (Table VI).Twenty of 23 patients had patent or stenosed less than 50%renal arteries at last follow-up with duplex ultrasound scan-ning. The mean pre-procedural serum creatinine concen-tration and creatinine clearance in the dialysis-dependentgroup was significantly impaired compared with the non-dialysis group. Thirteen patients showed worsening renal

Table IV. Proportional hazards and logistic regression mo

Patient characteristicsH

Associated with mortality*Creatinine deterioration within 3 moCreatinine deterioration after 3 moDialysisSystolic blood pressure � 140 mm Hg

Associated with morbidity*Female genderDiastolic blood pressure �90 mm Hg

Associated with immediate treatment failurefor hypertension*

Cerebrovascular diseaseSystolic blood pressure �140 mm Hg

Associated with long-term treatment failurefor hypertension*

HyperlipidemiaAssociated with immediate treatment failure

for creatinine*Preprocedure creatinine �2.5 mg/dL

Associated with long-term treatment failurefor creatinine*

Creatinine deterioration within 3 moTime to renal-related morbidity and

mortality*Hypertension treatment failure within 3 moCreatinine treatment failure within 3 mo

*P � .05.

function within 6 months after the procedure. Life tableanalysis at 3 years showed 20% mortality in the dialysisgroup and 9% mortality in the non-dialysis group (P �.001). Three deaths (13%) were attributed to renal causes,compared with non-dialysis renal mortality of 1.6%. Fivepatients with a pre-procedure serum creatinine concentra-tion less than 1.5 mg/dL receiving treatment of hyperten-sion alone and with ostial lesions became dialysis-depen-dent within 3 years of intervention. Three of these fivepatients died (60% considered renal-related). Two of thesepatients had immediate serum creatinine concentrationdeterioration within 3 months of the procedure. Two ofthese five patients (40%) required subsequent procedureson the ipsilateral vessel. All treated vessels remained patent.

DISCUSSION

Mortality. Renal artery stenosis is an independentpredictor of mortality. At 7 years, 73% of patients with renalartery stenosis are dead.13 The 5-year mortality in this studywas 25%, which is better than in patients with untreatedrenal artery stenosis, comparable to that in patients treatedwith surgical revascularization,14,15 and much better thanin other endoluminal studies.6-8 Immediate and long-termpost-procedure creatinine deterioration and dialysis depen-dency were associated with increased mortality in thisstudy. Elevated creatinine concentration affects survivalafter therapy for renal artery stenosis despite adequaterevascularization.5,7 Furthermore, increased survival maybe biased, on the basis of patient selection (ie, patientssuited for operative repair underwent endoluminal therapy,

rdso Odds ratio

95% Confidenceinterval

3 2.57–25.14 1.58–15.455 1.34–11.17 0.08–0.96

6.33 1.92–20.93.29 1.22–8.88

0.36 0.15–0.920.14 0.03–0.68

1 0.11–0.86

4.35 1.04–18.1

6 1.71–8.70

3 1.20–4.925.31–48.2

dels

azarati

8.04.93.80.2

0.3

3.8

2.416.0

JOURNAL OF VASCULAR SURGERYMarch 2004570 Sivamurthy et al

and typical risk factors for atherosclerosis-like coronaryartery disease did not affect survival), a demographicallymixed population with more women (compared with thestudy by Yutan et al6), a healthier cohort of patients, andpatients lost to follow-up. Despite these factors, regardlessof treatment indication, short-term and long-term treat-ment success for elevated creatinine positively affectssurvival.

Anatomic outcome and patency. Technical successof angioplasty or stenting at this institution is 98%. Thiscompares favorably with pooled data for angioplasty andstenting that show a success rate of 82% to 100%.15 Instudies of angioplasty, success is higher for nonostial lesions(72%-82%) than for ostial lesions (60%-62%). Ostial lesionsrepresent an extension of aortic atherosclerotic disease,making them more difficult to treat.16 In this study, 75% ofarteries were stented, and 75% of lesions were consideredostial; most stents (84%) were for ostial lesions. The use ofstents to correct ostial disease may explain the sustainedanatomic results.

Excellent primary patency and primary assisted patencyrates were also achieved at our institution (82% and 100% at5 years), with only one vessel occlusion at 5.5 years. A

Table V. Immediate clinical response

Indication n

Deteriora

n

Hypertension 126 0Elevated creatinine � 1.5 mg/dL 80 10Elevated Creatinine � 2.5 mg/dL 20 5Elevated creatinine 1.5–2.5 mg/dL 60 5

Indication Parameter n

All 146SBP (mm Hg)DBP (mm Hg)No. of medicationsCreatinine (mg/dL)

Hypertension 126SBP (mm Hg)DBP (mm Hg)No. of medicationsCreatinine (mg/dL)

Elevated creatinine 80SBP (mm Hg)DBP (mm Hg)No. of medicationsCreatinine (mg/dL)

Hypertension andelevatedcreatinine

62

SBP (mm Hg)DBP (mm Hg)No. of medicationsCreatinine (mg/dL)

DBP, Diastolic blood pressure; SBP, systolic blood pressure.*Number of patients and (percentages) of total treated patients for the resp†Values are mean � SD.

recent study of endoluminal stents placed to treat renalartery stenosis demonstrated a primary patency rate of 77%at an average of 18 months of follow-up.17 A comparison ofangioplasty and surgery shows primary recurrent stenosis–free patency rates of 17% to 75% and 81% to 96%, respec-tively, at 4 years.8,18 Comparison of our study results withthe natural history of medically managed renal artery ste-nosis3,19,20 may suggest that percutaneous therapy is nobetter than medical management, on the basis of patencydata. The natural rate of occlusion is 5% per year, and therisk for occlusion is greater in vessels with higher stenoses.A problem with natural history studies is that candidatesnot considered for intervention were included, and therewas significant crossover of patients requiring intervention.In addition, patients in whom the role of early renal revas-cularization was uncertain were included. In our study,patients had an average stenosis of 80%, that is, high-gradestenoses more likely to occlude. Furthermore, the 18%failed primary patency rate at 5 years represents vesselstreated for symptomatic recurrent stenosis, because therewere no occlusions within 5 years. Patency was improvedwith secondary interventions to 100%, better than thenatural history of this disease. This emphasizes the need for

Response*

No effect improved cured

n % n % n %

61 48 49 39 16 1360 75 9 11 1 113 65 2 10 0 047 78 7 12 1 2

Preprocedure† Postprocedure† P

176 � 30 161 � 25 �.0188 � 18 83 � 15 �.01

2.9 � 1.0 2.8 � 1.2 � .071.8 � 1.2 1.7 � 0.9 � .87

181 � 28 159 � 23 �.0190 � 17 84 � 14 �.01

3.0 � 1.0 2.8 � 1.2 � .101.7 � 1.2 1.6 � 0.9 � .39

168 � 29 157 � 25 � .187 � 18 84 � 15 � .03

3.1 � 1.0 2.8 � 1.2 � .92.4 � 1.3 2.2 � 1.0 � .5

171 � 27 155 � 24 � .0390 � 17 86 � 14 � .01

3.1 � 1.0 2.8 � 1.3 � .352.4 � 1.4 2.2 � 1.0 � .42

subgroup.

tion

%

013258

ective

JOURNAL OF VASCULAR SURGERYVolume 39, Number 3 Sivamurthy et al 571

Top, Life table analysis of the entire study population (black circles), patients with hypertension (gray circles), and patients with renaldysfunction (triangles). Middle, Cumulative primary patency (gray circles), primary assisted patency (black circles), and primary recurrentstenosis–free (triangles) intervals for all vessels. Bottom, Life table analysis of the likelihood of continued deterioration in blood pressure(black circles) and renal function (gray circles) after renal angioplasty with or without stenting. The number of patients or vessels in eachinterval is given in the table to the left of the corresponding graph. Values represent mean � SE �10% for all data points.

JOURNAL OF VASCULAR SURGERYMarch 2004572 Sivamurthy et al

vessel surveillance and the importance of secondary endolu-minal procedures to maintain vessel patency after the pri-mary intervention.

Recurrent stenosis for angioplasty of atheroscleroticrenal artery stenosis ranges from 10% to 47%, which isattributed to the elastic recoil and rigidity of these lesions,potential for arterial dissection, and inability to cope withostial plaque.15 The use of stents decreases the rate ofrecurrent stenosis to 11% to 23% at 1 year, consistent withour 5-year recurrent stenosis rate.15,21 No factors wereidentified that contributed to recurrent stenosis, includingvessel size, placement of a stent, or the presence of cardio-vascular risk factors. Although recurrent stenosis could beadequately treated to maintain vessel patency, it did notaffect functional outcomes.

Functional outcome. Maintenance of clinical benefitfrom percutaneous therapy for hypertension and renal dys-function was mixed and not so effective as surgery.14 Amodest 52% of patients who received treatment of hyper-tension had immediate improvement, and no patients weremade worse by the treatment. This result is slightly lowerthan the 68% obtained by Yutan et al.6 Recent studiessuggest that the use of endoluminal therapy to cure hyper-tension from atherosclerotic renal artery stenosis israre.7,9,21,22 Usually there is an improvement in bloodpressure and a reduction in the number of medications.22

Webster et al22 compared angioplasty with best medical

Table VI. Comparison of patients requiring or not requiri

Patient characteristics

Dia

n

Total 23Male 7Female 16Indication

Hypertension alone 5Elevated creatinine alone 2Hypertension and elevated creatinine 16

Stent placement 20Complications 2†

Hypertension alone 0Elevated creatinine alone 0Hypertension and elevated creatinine 2

Total mortality at 3 y 7Renal-related mortality 3

Hypertension alone 2Elevated creatinine alone 0Hypertension and elevated creatinine 1

Patient characteristics Mean � S

Pre-procedure creatinine (mg/dL) 3.1 � 2.Pre-procedure creatinine clearance (mL/min) 32 � 22Freedom from dialysis, renal-related death, or

�20% increase in creatinine (y)0.9 � 1.

Duration of blood pressure benefit (y) 2.9 � 2.Average follow-up (y) 2.7 � 2.

*Percent of total number of patients.†Peripheral embolization and stent dislodgement to iliac arteries.

therapy, and found no difference between treatments forhypertension in unilateral disease. However, in bilateraldisease angioplasty can decrease blood pressure and reducethe number of medications required. Results of this studymay be explained by the 82% of bilateral disease that wastreated. Furthermore, reduction in systolic and diastolicblood pressure was statistically significant without a statis-tically significant reduction in medications. However, it wasdifficult to assess medications, because many medicationsused in this patient population have dual indications. Bloodpressure response had no effect on follow-up cardiovascularmorbidity and patient mortality. Although systolic bloodpressure was statistically lowered with treatment, by itself itmay not have been enough to change the category ofresponse (Table V). However, immediate improvement ofhypertension from percutaneous therapy positively affectedrenal function and freedom from renal-related mortality(Table IV). Extended analysis of patients in this studyshows a modest number of patients maintained the clinicalimprovement in hypertension. Steinbach et al23 and Yutanet al6 reported long-term improvement of hypertension at72% and 52%, respectively.

Functional outcome in patients with renal impairmentwas not so promising. Creatinine concentration was stabi-lized or decreased in 87% of patients in the first 3 months.However, these improvements were not maintained duringextended follow-up (45% at 5 years). When stratified by

ialysis post-procedure

No dialysis

P% n %

15.8* 123 84.2*30.4 59 48.069.6 64 52.0

21.7 59 48.08.7 18 14.6

69.6 46 37.4117

8.7 31 25.21816

20.0 12 9.0 �.0113 2 1.6

011

Range Mean � SD Range P

0.9–9.1 1.6 � 0.7 0.6–4.3 �.013–82 49 � 26 9–164 �.010–4 1.7 � 1.8 0–8 �.01

0–6 1.9 � 1.9 0–8 �.030–6.6 1.6 � 2.0 0–10 �.01

ng d

lysis

D

2

0

00

JOURNAL OF VASCULAR SURGERYVolume 39, Number 3 Sivamurthy et al 573

serum creatinine concentration, patients with an elevatedcreatinine concentration (�2.5 mg/dL) had significantlyincreased mortality and a poor rate of retained clinicalbenefit. Even with such stratification, long- term functionaloutcome was poor, with no subgroup demonstrating im-provement. However, patients with immediate stabiliza-tion or improvement in serum creatinine concentration hada greater rate of freedom from dialysis and renal-relatedmortality. When compared with surgery, the immediateimprovement in creatinine concentration (12%) and thelong-term maintenance are poor. These observations arelikely secondary to the definition of the outcome catego-ries, the limitation of serum creatinine in determining poorrenal function, our standard for nonfunctioning kidneys,limitations of the procedure, and patient selection. In 75%of patients serum creatinine concentration stabilized imme-diately after therapy. This category of patients needs to bebetter defined and optimized, and includes patients whomay have had divergent outcomes over extended follow-up. Although not specifically investigated in this study, thecategory of renal dysfunction contributes to clinical out-come after surgical revascularization.24 Thirty-six percentof patients with chronic renal failure and 62% of patientswith rapidly declining renal function had improvement inrenal function after revascularization at 19-month follow-up.25 We acknowledge the limitation in measuring serumcreatinine concentration to assess the early response totreatment. In patients who had a modest preoperativecreatinine concentration of less than 2.0 mg/dL, earlyimprovements in the serum creatinine level were not evi-dent despite measurable improvements in glomerular filtra-tion rate.26 Serum creatinine concentration may not be thebest outcome measure to determine treatment response,and may have been too late in identifying those patientswho had minimal end organ disease. With rising creatinineconcentration, renal function has declined significantly.Revascularization before creatinine concentration rises andparenchymal mass is lost may be a better approach.12 Lateoutcomes with surgery and percutaneous therapy are betterwhen pre-procedure renal function is normal, as noted inthis study.12,23 No factors at duplex ultrasound scanningpredicted response to therapy in this study. Furthermore,this study may have been more aggressive with “nonfunc-tioning” kidneys, because some of the criteria used todefine nonfunctioning kidneys do not correlate with recov-ery of function after open repair. Gill-Leertouwer et al24

suggests that in patients with renal impairment lateraliza-tion to the affected kidney on a renal scintigram is apredictor of clinical success. More “nonfunctioning” kid-neys may have been treated with our criteria.

Limitations of the procedure include inability to treatdistal disease in the kidney, and atheroemboli. Of concernin this study is that 23 patients (16%) became dialysis-dependent after the procedure. Furthermore, this subset ofpatients had a 20% mortality rate, compared with 9% in thenon-dialysis group, at 3 years. More disturbing is that fivepatients who received treatment of hypertension alone,with normal serum creatinine concentration, kidney size,

resistive index, and creatinine clearance, became dialysis-dependent. The increased number of patients requiringdialysis post-procedure in this study emphasizes the riskassociated with endovascular renal revascularization andthe potential need for distal protection devices to improvethe outcome of endoluminal therapy for renal artery steno-sis.27 Patient selection may have also contributed to theresults. Early in our experience we were more liberal in ourtreatment, but have since become more conservative. Anegative selection bias from treating severe disease, ostialdisease, and bilateral disease may have influenced out-comes. These functional results indicate that it is essentialto develop better criteria to identify those at risk for treat-ment failure and better outcome measures of renal failure.

This study does not advocate the liberal use of catheterintervention for renal artery stenosis. Despite the excellenttechnical outcomes, this study points out the limitations ofintervention and the need for proper patient selection.There was a higher complication rate than in other stud-ies.5,6 Although most complications were minor, four pa-tients required surgery to address them. Our techniqueimproved with experience and time. In addition, 16% ofpatients became dialysis-dependent. It is unclear whetherthese patients would have required dialysis without treat-ment. However, the five patients with normal creatinineconcentration who became dialysis-dependent emphasizethe risk of this procedure. We may have been too aggressivein our treatment protocol. Further assessment of functionaloutcomes was not so promising. This study may haveunderestimated the extent of renal disease and end organinjury by using serum creatinine concentration, especiallyin patients with stable renal function after the procedure.Future studies should consider better criteria to determinerenal dysfunction and categorize outcomes. The goalshould be to treat only symptomatic renal artery stenosiswith objective evidence of reversible end organ disease.Reviewing natural history data for renal artery stenosis,progression to occlusion occurs at a modest rate, especiallywith high-grade stenosis. Furthermore, patients who re-ceived medical management had decreasing kidney sizeover time, which emphasizes the need for invasive therapy.3

However, this study shows that all patients with symptom-atic stenosis do not benefit from therapy. Medical manage-ment may still be appropriate in a subset of patients.

In conclusion, this study shows that endoluminal inter-vention to treat symptomatic atherosclerotic renal arterystenosis is technically successful, with excellent anatomicresults. There are excellent patency and modest recurrentstenosis, which can be treated with secondary interven-tions. The number of patients unexpectedly requiring dial-ysis after percutaneous therapy and the number of proce-dural complications cautions against aggressiveprophylactic therapy. Despite these limitations, patientswith immediate treatment success were more likely to havegreater freedom from renal morbidity and mortality. Insummary, endovascular interventions may enhance the careof the patient with hypertension, but does not appear topreserve long- term renal function in the patient with

JOURNAL OF VASCULAR SURGERYMarch 2004574 Sivamurthy et al

chronic renal impairment. Further analysis and randomizedstudies will be required to better define these disappointingoutcomes in chronic renal insufficiency and to determinethe appropriate time frame in which to pursue revascular-ization.

REFERENCES

1. Caps MT, Zierler RE, Polissar NL, Bergelin RO, Beach KW, Cantwell-Gab K, et al. Risk of atrophy in kidneys with atherosclerotic renal arterystenosis. Kidney Int 1998;53:735-42.

2. Caps MT, Perissinotto C, Zierler RE, Polissar NL, Bergelin RO, TullisMJ, et al. Prospective study of atherosclerotic disease progression in therenal artery. Circulation 1998;98:2866-72.

3. Zierler RE, Bergelin RO, Isaacson JA, Strandness DE Jr. Natural historyof atherosclerotic renal artery stenosis: a prospective study with duplexultrasonography. J Vasc Surg 1994;19:250-8.

4. Cambria RP, Brewster DC, L’Italien GJ, et al. The durability of differ-ent reconstructive techniques for atherosclerotic renal artery disease. JVasc Surg 1994;20:76-87.

5. Lederman RJ, Mendelsohn FO, Santos R, Phillips HR, Stack RS,Crowley JJ. Primary renal artery stenting: characteristics and outcomesafter 363 procedures. Am Heart J 2001;142:314-23.

6. Yutan E, Glickerman DJ, Caps MT, Hatsukami T, Harley JD, KohlerTR, et al. Percutaneous transluminal revascularization for renal arterystenosis: Veterans Affairs Puget Sound Health Care System experience.J Vasc Surg 2001;34:685-93.

7. Dorros G, Jaff M, Mathiak L, Dorros II, Lowe A, Murphy K, et al.Four-year follow up of Palmaz-Schatz stent revascularization as treat-ment of atherosclerotic renal artery stenosis. Circulation1998;98:642-7.

8. Weibull H, Bergqvist D, Bergentz SE, Jonsson K, Hulthen L, ManhemP. Percutaneous transluminal renal angioplasty versus surgical recon-struction of atherosclerotic renal artery stenosis: a prospective random-ized study. J Vasc Surg 1993;18:841-52.

9. Van Jaarsveld BC, Krijnen P, Pieterman H, Derkx FH, Deinum J,Postma CT, et al. The effect of balloon angioplasty on hypertension inatherosclerotic renal-artery stenosis. N EngL J Med 2000;342:1007-14.

10. Harden PN, MacLeod MJ, Rodger RSC, Baxter GM, Connell JM,Dominiczak AF, et al. Effect of renal artery stenting on progression ofrenovascular failure. Lancet 1997;349:1133-6.

11. Rundback JH, Sacks D, Kent KC, Cooper C, Jones D, Murphy T, et al.Guidelines for the reporting of renal artery revascularization in clinicaltrials. J Vasc Interv Radiol 2002;13:959-74.

12. Cambria RP, Brewster DC, L’Italien GJ, Gertler JP, Abbott WM,LaMuraglia GM, et al. Renal artery reconstruction for the preservationof renal function. J Vasc Surg 1996;24:371-82.

13. Zierler RE, Bergelin RO, Davidson RC, Cantwell-Gab K, Polissar NL,Strandness DE. A prospective study of disease progression in patientswith atherosclerotic renal artery stenosis. Am J Hypertens 1996;9:1055-61.

14. Cherr GS, Hansen KJ, Craven TE, Edwards MS, Ligush J, Levy PJ, etal. Surgical management of atherosclerotic renovascular disease. J VascSurg 2002;35:236-45.

15. Kidney D, Deutch LS. The indications and results of percutaneoustransluminal angioplasty and stenting in renal artery stenosis. SeminVasc Surg 1996;9:188-97.

16. Bonelli FS, McKusick MA, Textor SL, Kos PB, Stanson AW, JohnsonCM, et al. Renal artery angioplasty: technical results and clinical out-come in 320 patients. Mayo Clin Proc 1995;70:1041-52.

17. Siablis D, Liatsikus EN, Kalogeropoulous C, Karnahatides D, Tsotu I,Periminus P, et al. Metal stents for the management of atheroscleroticrenovascular disease. J Endourol 2001;15:993-6.

18. Erdoes LD, Bermans S, Hunter GC, Mills JL. Comparative analysis ofpercutaneous transluminal angioplasty and operation for renal revascu-larization. Am J Kidney Dis 1996;27:496-503.

19. Schrieber MJ, Pohl MA, Novick AC. The natural history of atheroscle-rotic and fibrous renal artery disease. Urol Clin North Am 1984;11:383-92.

20. Tollefson DF, Ernst CB. Natural history of atherosclerotic renal arterystenosis associated with aortic disease. J Vasc Surg 1991;14:327-31.

21. Dorros G, Jaff M, Jain A, Dufeh C, Mathiah L. Follow up of primaryPalmaz-Schatz stent placement for atherosclerotic renal artery stenosis.Am J Cardiol 1995;75:1051-5.

22. Webster J, Marshall F, Abdalla M, Dominiczak A, Edwards R, Isles CG,et al. Randomized comparison of percutaneous angioplasty versus con-tinued medical therapy for hypertensive patients with atheromatousrenal artery stenosis. J Hum Hypertens 1998;12:329-35.

23. Steinbach F, Novick AC, Campbell S, Dykstra D. Long-term survivalafter surgical revascularization for atherosclerotic renal artery disease.J Urol 1997;158:36-41.

24. Gill-Leertouwer TC, Gussenhoven EJ, Bosch JL, Deinum J, van Over-hagen H, Derkx FH, et al. Predictors for clinical success at one yearfollowing renal artery stent placement. J Endovasc Ther 2002;9:495-502.

25. Cognet F, Garcier JM, Dranssart M, Defraissinette B, Cercvel JP, RavelA, et al. Percutaneous transluminal angioplasty in atheroma with renalfailure: long term outcome in 99 patients. Eur Radiol 2001;11:2524-30.

26. Dean RH, Englund R, Dupont WD, Meacham TW, Plummer WD,Pierre R, et al. Retrieval of renal function by revascularization. Ann Surg1985;202:367-74.

27. Henry M, Klonaris C, Henry I, Tzetanov K, Le Borgne E, Foliguet B,et al. Protected renal stenting with the PercuSurge GuardWire device: apilot study. J Endovasc Ther 2001;8:222-37.

Submitted Mar 6, 2003; accepted Sep 12, 2003.